info-vax (06/23/82)
>From reece@NADC Wed Jun 23 13:20:24 1982
I have a vax 11/780 running 4.1BSD UNIX with two Systems Indus-
tries RM05 type drives with their 9400 controller hooked to the
SBI. Almost daily I have a problem where one or the other drives
is "hung" and has to be cleared by hitting the reset on the 9400.
SI techs have replaced everything north of the physical drives
with no positive results. The latest theory is a firmware problem
in the 9400. They say other customers have had similar problems.
The register dump always shows the last command to be "search".
One fix they came up with is to delete all the search commands
from the op system. Does anyone out there identify with this
problem and (hopefully) know what the fix is? I've been reset-
ting the system daily for about 4 months.
Jim Reece
-------info-vax (06/25/82)
>From mo@LBL-UNIX Fri Jun 25 14:11:39 1982
I have had a similar configuration for over a year and a half now
and it has performed flawlessly. Is the system newly installed, or
is this a recent development? Are they radial or daisy chain??
What boards have they changed in the 9400 itself? Drive boards,
motherboard? What rev microcode do you have??
-Mikeinfo-vax (07/04/82)
>From decvax!harpo!ihnss!ihps3!stolaf!sft@Berkeley Sat Jul 3 19:37:11 1982
We have had similar problems with SI 9400 controllers on an 11/60 running
RSTS and an 11/70 running UNIX V7. Our VAX/VMS system seems to be immune.
There is indeed a firmware problem with overlapped seeks in 9400s, and it
sounds to me like you've found it. Our fix in V7 (and 2.8 BSD) UNIX is to
delete the code in hpstart() between the labels "search:" and "done:".
This disables overlapped seeks and makes the system run, although a little
slower. Sorry I can't help you with 4.1BSD--we're still waiting for our
VAX license and I don't have source code.
On the other hand, you might have a hardware problem. One of our controllers
(on the 11/60) had a tendency to go south, usually between midnight and
7 AM. The controller would indicate strange things (including drive off-line)
in the drive status registers and would refuse to accept any commands.
Replacing the main logic board fixed that. I'd bet that your problem is
firmware, though.
Steve Tarr
....!ihnss!ihps3!stolaf!sft